Summary Macro-autophagy is the intracellular stress-response pathway by which the cell packages portions of the cytosol for delivery into the lysosome. This packaging is carried out by the de novo formation of a new organelle called the autophagosome that grows and encapsulates cytosolic material for eventual lysosomal degradation. How autophagosomes form, including especially from where the autophagosomes extract lipid in order to expand their membranes, has been a core problem in the field for over 50 years. Two competing models have emerged and suggest that either the autophagosome simply grows out of a pre-existing compartment (like the ER) or the autophagosome forms from the continued fusion of individual vesicles recruited from many different sites in the cell. Furthermore, lipid biogenesis pathways are intimately associated with autophagy, suggesting that one or more organelles involved in the production of lipids might also be tightly associated to the growth of the autophagosome. Temporal studies established over a decade ago that the autophagy protein ATG2, is needed during membrane expansion, but how ATG2 facilitates membrane growth has remained elusive. Our preliminary results now demonstrate that ATG2 is a member of a novel lipid-transport family of proteins and suggest a third model for membrane expansion; the bulk delivery of lipid from organelles through protein- mediated contact sites. Indeed, ATG2 binds up to 20 lipids at once, an order of magnitude more than virtually any other lipid transport protein, and thus has the capacity to move a lot of lipid during biogenesis. We show that in cells, ATG2 accumulates at an interface between autophagosomes and the ER, strongly suggesting this organelle-organelle contact site might be the location of lipid transfer. In addition, we have developed gene- edited knockouts of each of the ATG2 proteins in humans and discovered that in the absence of ATG2, not only do autophagosomes not expand, but hundreds of vesicles collect at the site of autophagosome biogenesis. This surprising observation suggests that vesicle-mediated delivery of membrane might also be essential and that the fusion of these vesicles is specifically disrupted when ATG2-mediated lipid transport is absent. With this proposal, we expect to describe how ATG2 works with proteins on both the autophagosome and the ER to drive lipid flow. Likely, this will involve proteins needed to stabilize organelle-organelle contact sites and may also involve proteins sensing or regulating lipid production in the ER. We will then establish how this lipid flow is related to the recruitment and utilization of trafficking vesicles to describe to support autophagosome growth.